5 May
2008
5 May
'08
1:18 p.m.
Hi all,
I recently obtained an old PDP-8/L in reasonable shape (why does this sound
like an AA meeting? :-)
Anyway, it doesn't work. So I'm happy debugging. The supplies on the
back plane are not right so I decided to pull the main supply.
I finally broken down and ordered an ESR meter. Is it smart to check
all the caps?
I've never used an ESR meter but it seems like a good idea to try it on
all the big caps just to see what sort of shape they are in. I assume I
need to open the little caps which are in parallel to get a good reading
(my experience, limited, is that the little axial electrolytics tend to
open/short much more often than the big cans).
One of the unregulated supplies (-6) coming off the main transformer was
a volt low when loaded and I thought I'd check the caps and the diodes.
Unloaded it shows -33v (but looking at the schematic, that may be
actually correct. Loaded the +5 was low (like 4.5), unloaded it's
perfect (5.01).
The lights do light, and run sets the run light, stop stops but load
address does not. Worse, it seems to toggle the MA, which is ringing a
bell about a common fault. I figured I'd get the supplies all cleaned
up first...
-brad
5 May
5 May
1:46 p.m.
Brad wrote:
I finally broken down and ordered an ESR meter. Is it
smart to check
all the caps?
You don't really need an ESR meter for that. A lab power supply
(i.e., adjustable current limit) will work fine. For a DC power supply,
the main things you look for are:
1) Is the capacitor excessively leaky at the operating voltage?
It shouldn't pass much DC. Set the PS voltage to the operating
voltage of the capacitor, and set the current limit low (say
tens of mA). Connect the capacitor and see what the voltage does.
If the capacitor is good, the PS will start in current limiting,
and the voltage will go up linearly as the capacitor charges.
Eventually the voltage will hit the supply setting, and the
capacitor should draw almost no current at this point.
If the voltage gets stuck at some point, that means the capacitor
is leaky. It may just need reforming; if so, leaving it on the PS
for a while may do the trick.
2) Is the capacitance within spec? Measure how long it takes to
charge. If you're using constant voltage, the rate of charge
exhibits exponential decay. If you use constant current, it is
linear.
Eric
2:49 p.m.
On Mon, May 05, 2008 at 04:18:57PM -0400, Brad Parker wrote:
Hi all,
I recently obtained an old PDP-8/L in reasonable shape...
Nice. That was my first model of -8. I still marvel at what
they packed into that package (and how far you can expand it
with an extra box or two).
Anyway, it doesn't work. So I'm happy
debugging. The supplies on the
back plane are not right so I decided to pull the main supply.
.
.
.
One of the unregulated supplies (-6) coming off the
main transformer was
a volt low when loaded and I thought I'd check the caps and the diodes.
Unloaded it shows -33v (but looking at the schematic, that may be
actually correct. Loaded the +5 was low (like 4.5), unloaded it's
perfect (5.01).
You might characterize at what load it droops. I had an arcade game
that arrived dead because the regulator transistor was faulty - the
supply did put out +5.0 V with no load, but more than a few mA of
load and the supply just wouldn't put out anything close to +5V.
OTOH, you mention 4.5V, which is at the bottom end of the spectrum
for TTL. Obviously closer to +5.0 is good, but what you are reporting
below might or might not be power related.
The lights do light, and run sets the run light, stop
stops but load
address does not. Worse, it seems to toggle the MA, which is ringing a
bell about a common fault. I figured I'd get the supplies all cleaned
up first...
Obviously it's good to be confident about your power, but if you get that
issue addressed and still see odd symptoms, I'd grab the printset from
bitsavers and keep a careful eye to any 7474 and 7440 chips that are
in a path that could affect the behavior of whatever front-panel activity
you are trying. In my experience with repairing several -8/Ls and an
-8/i, those two chip types fail at a remarkably higher rate than all the
other ICs combined.
For testing simple boards (M111, M113, M117, M216), I have rigged up
a squeeze-on 16-pin DIP clip to a hand-held IC tester. There's enough
power from the tester battery to power 3-4 TTL ICs, and none of the
signals on the simple boards interact on the board. It only takes a
few seconds to check all the 7474s on an M216 that way. It may not
find subtle chip problems, but in practice, I've frequently found a
chip or two that's wonky. Unfortunately, this trick won't work on
the M220 CPU register modules, nor I/O like the M706/M707 console
interface boards, but as easy as it is to check 80% of the boards
in the machine and eliminate them from your scoping, I think it's
quite a valuable technique.
The other issue that comes up is backplane oxidation. What has done
wonders for me in the past is to take manila folders, cut them into
strips the width of the card fingers, fold into a 2-thickness layer,
then wet down with 1,2-propanol (isopropyl) and insert/remove a few
times. You will probably see dark grey streaks on the paper the first
couple of times you do it.
Power is a great place to start, though. I can't really give much
advice there; I've never had to delve into the PSU of an -8/L.
-ethan
--
Ethan Dicks, A-333-S Current South Pole Weather at 5-May-2008 at 21:30 Z
South Pole Station
PSC 468 Box 400 Temp -71.1 F (-57.3 C) Windchill -97.9 F (-72.2 C)
APO AP 96598 Wind 5.7 kts Grid 62 Barometer 697.4 mb (9979 ft )
Ethan.Dicks at usap.gov http://penguincentral.com/penguincentral.html
8:13 p.m.
[excellent advice, btw - thanks!]
but this one I was wondering about, and I like your technique:
Ethan Dicks wrote:
...
The other issue that comes up is backplane oxidation.
What has done
wonders for me in the past is to take manila folders, cut them into
strips the width of the card fingers, fold into a 2-thickness layer,
then wet down with 1,2-propanol (isopropyl) and insert/remove a few
times. You will probably see dark grey streaks on the paper the first
couple of times you do it.
(I owe you a few pints!)
-brad
6612
days inactive
6613
days old
3 comments
3 participants
participants (3)
-
brad＠heeltoe.com -
eric＠brouhaha.com -
ethan.dicks＠usap.gov